Comment by danpalmer

I think that's one interpretation. Another is that proving the privacy and data protection aspect takes longer, regardless of whether the correct work has been done. Another interpretation is that it's not about data protection or privacy, but about AI regulation (even prospective regulation), and that they want to be cautious about launches in regions where regulators are taking a keen interest.

I'm biased here, but based on my general engineering experience I wouldn't expect it to be about privacy/data protection. As a user I think things like Wipeout/Takeout, which have existed for a long time, show that Google takes this stuff seriously.